Automated and manual leukocyte differential counts in healthy equines

The speed and convenience provided by automated cell counters are reasons for the increasing use of this technology in veterinary practice, with impedance methodology being one of the most commonly used in Brazil. In this regard, the objective of this study was to compare the automated differential leukocyte count obtained by impedance with the manual count obtained from the evaluation of blood smears in healthy horses. For this purpose, the automated differential leukocyte count using a veterinary cell counter (Abx Micros ESV 60, Horiba) was compared to the manual count using optical microscopy in 545 blood samples from healthy horses. Paired t-tests or Wilcoxon tests, Deming regression, Pearson or Spearman correlations, and Bland-Altman analysis were used, considering differences significant when p<0.05. The automated method showed higher concentrations of segmented neutrophils and monocytes, and lower concentrations of lymphocytes and eosinophils compared to the manual method. The mean error of the automated count was -10.26% for segmented neutrophils, 11.04% for lymphocytes, -41.39% for monocytes, and -10.84% for eosinophils, ranging from -200% to 161.4% depending on the cell type. There was a significant correlation between the methodologies only for the counts of segmented neutrophils, lymphocytes, and eosinophils. In conclusion, the manual differential leukocyte count in healthy horses cannot be replaced by the automated method, and the analysis of blood smears remains a fundamental tool for the proper interpretation of the leukogram in horses.